Transcript GENI: Global Environment for Networking Innovations

GENI: Global Environment for
Networking Innovations
To Reinvent the Internet
Guru Parulkar
CISE/NSF
[email protected]
Internet Has Been A
Transformative Infrastructure
Top 20 Engineering Innovations of
20th Century
A Century of Innovations
(National Academy of Engineering)
State of Internet
“… in the thirty-odd years since its invention, new uses
and abuses, …, are pushing the Internet into realms
that its original design neither anticipated nor easily
accommodates.”
“Freezing forevermore the current architecture would
be bad enough, but in fact the situation is deteriorating.
These architectural barnacles—unsightly outcroppings
that have affixed themselves to an unmoving
architecture— may serve a valuable short-term
purpose, but significantly impair the long-term flexibility,
reliability, security, and manageability of the Internet.”
Overcoming Barriers to Disruptive Innovation in Networking, NSF Workshp Report, 05.
3
GENI Initiative: Goals
• Invent innovative internet architecture(s) and
distributed system capabilities -- go beyond Internet
• Enable seamless conception-to-deployment process
– Facility for experimentation at scale with apps and users
• Work with broader community
– Academic and industry communities
– Other US and international agencies
4
Outline
• GENI research rationale
• GENI facility rationale and design
• Opportunities for participation
• Example architecture themes and systems
5
The Internet Hourglass
Applications
Kazaa VoIP Mail
TCP
News
Video
Audio
IM
U Tube
Everything
Transport protocols
on IP
SCTP
UDP
ICMP
IP
Ethernet 802.11
IP on
Power lines ATM Optical
everything
Satellite Bluetooth
Link technologies
Modified John Doyle Slide 6
Internet and WEB Hourglass
Kazaa VoIP Mail
Applications
Everything
Newson Video
WEB Audio
IM
U Tube
HTTP
Ossification
Transport protocols
TCP
SIP
UDP
RTP
Continued
Innovations
IP
Ethernet 802.11
on
Power lines IP
ATM
everything
Optical
Satellite Bluetooth
Modified John Doyle Slide 7
Internet Architecture Concerns
• Security and robustness
• Control and management
• Addressing, naming & (inter-domain) routing
• End-to-end principle vs in-network processing
• Mobility of hosts and networks
• Economic viability of different stakeholders
8
Mobile Wireless Trends
• 2B+ cell phones
• 400M+ cell phones with Internet capability -- rising rapidly
• New data devices (blackberry, PDA, iPod) and services
• 240M vehicles on the road -- to get network connectivity
• Mobile wireless to dominate future computing
and communication
• Current Internet not designed for mobile
wireless
9
Embedded Physical World
New Machines
New Environments
New Applications
New Scale
Billion to trillion devices!
10
Networking the Physical World
Internet
Internet architecture not designed to network the
physical world
11
Photonics Integration Trends
Source: Vinod Khosla
Source: Dan Blumenthal
10ps Delay using deep-etched waveguide
Dual SGDBR Signal
Booster SOAs
Label Rewrite EAM
SGDBR
Flared Input Pre- 1mm MZI SOAs
Input Signal Blanking EAM Tunable Laser
amp SOAs
Almost Free OEO!!
1.6 Tbps on a chip: 40 x 40
Packet Switching
Buffering in Optical Domain
High capacity “dynamic” optical networks a certainty
Current Internet architecture not designed to take advantage of it
12
Distributed System Capabilities
• Information dissemination
– Standard and streaming media; stored and real time
• Management and sharing of personal information
– Balance of privacy and accountability
• Networked embedded systems
– Networked robotics, power grid, and smart spaces
• Location aware services: human and object
• Identity management: human and object
• And many more …
13
Digital Living 2010
Tomorrow’s users will be surrounded by pervasive devices,
embedded sensors and systems… all connected to the Internet.
User
User
Communications
Games
Photography
Inventory/Sales
tracking
Entertainment Systems
Health/Medical
Home Computer
Home Appliances
Banking
and
Commerce
User
Surveillance and Security
(at home, work, or in public)
PDA
Telephone
Car
Building Automation
User
Thanks to David Kotz at Dartmouth 14
Network Centric Critical Infrastructures
Essential Utilities
Transportation
Internet Architecture
is NOT robust enoughTelecommunications
to support these Banking & Finance
15
Future Internet?
Distributed Systems and Services?
Network and Protocol Architectures?
New Paradigms?
Internet
Arch Limitations
Erosion
Enabling
Technologies
Network
Capabilities
Applications &
User Requirements
Need a clean-slate approach
16
GENI: Research Programs
Broad but goal oriented programs: New internet architectures and
distributed system capabilities
NeTS
FIND: Future Internet Design
CyberTrust
Clean-slate secured network arch
CSR
New distributed system capabilities
CCF
SING
CRI
Network infrastructure for arch research
MRI
Network instruments for arch research
IIS Program ??
17
Fundamental Change in Mindset
Needed
• Don’t ask how a new
technology/application fits within the
existing Internet architecture
• Think how a new
technology/network/application can shape
the Future Internet in realizing its
potential?
18
Outline
• GENI research rationale
• GENI facility rationale and design
• Opportunities for participation
• Example architecture themes and systems
19
Conception-to-Deployment
Case for GENI Facility
Maturity
Need for Large experimental
facility/infrastructure
Shared
Deployed
Infrastructure
This chasm represents a major
barrier to impact real world
Small Scale
Testbeds
Research
Prototypes
Foundations
Research
Funded by CISE Programs
Time
20
Facility Goals
Enable exploration of new network architectures,
mechanisms, and distributed system capabilities
A shared facility that allows
• Concurrent exploration of a broad range of experimental
networks and distributed services
• Interconnection among experimental networks & the
commodity Internet
• Users and applications able to “opt-in”
• Observation, measurement, and recording of outcomes
Help develop stronger scientific base
21
Facility Design: Key Concepts
Sensor Network
Edge Site
Mobile Wireless Network
Federated Facilities
Slicing, Virtualization, Programmability
22
Facility Architecture: How do
Parts Fit Together?
Infrastructure services
RDS PS
CS
MS
LS
. . . and others . . .
Slice Manager
Core
Resource Controller
Auditing Archive
node
control
sensor
data
CM
CM
CM
Node substrate
Node substrate
Node substrate
Components
23
Scope of Research & Experiments
Applications
Distributed Systems and Services
Dist Sys X
Dist Sys Y
Dist Sys Z
E2E Architectures and Components
Arch Comp
X
Arch Comp
Y
Applications
Service
C
Arch
Arch B C
Service-oriented
Applications
Service
Arch B
A B
Arch
Service-oriented
Arch A
Service A
Applications
Arch Comp
Z
GENI Facility Substrate: Sharable Physical Infrastructure
Backbones
Campus/Access
Backbone
Edge Networks
Combination of All optical transport and
Packet service networks
Numerous Wired, Mobile Wireless, and Sensor Networks
Campus/Access
Backbone
Edge Networks
24
GENI Status
• Research programs
– Several launched and others getting formulated
• Facility
– Facility definition proceeding as per MREFC process
– The big funding expected in FY10
25
GENI Status: Another Look
• Stage is set to focus on “reinventing the Internet”
– Change of mindset
– Priority setting and resource commitment
– High level research direction and facility definition
• Fun begins: exciting opportunities/challenges
– Identify specific research directions
– Explore innovative ideas and rewarding research
– Develop technologies and systems
Help change the world -- one more time
26
Outline
• GENI research rationale
• GENI facility rationale and design
• Opportunities for participation
• Example architecture themes and systems
27
Opportunities for Participation
• Submit targeted proposals to programs
– FIND, Cyber Trust, CSR, …
• Participate in GENI definition
– Town hall meetings
– Discussion lists
– Working groups and planning group
• Plan to build and use components of GENI
– CRI and MRI programs
– Work with GENI Project Office
• Plan to use GENI
– Propose and prototype exciting new architectures
28
Stanford’s Opportunity to Lead
• Shared vision: clean slate program
• Depth and breadth of expertise
• Availability of resources
– University, industry, state, federal agencies
• Reinventing Internet is “Stanford-hard” challenge
29
Outline
• GENI research rationale
• GENI facility rationale and design
• Opportunities for participation
• Example architecture themes and systems
30
Architectural Themes
• Innovation oriented architecture (IOA)
• Service oriented architecture
• Technology driven cross layer architecture
31
IOA Theme Rationale
• Two extremes have dominated telecom so far
– Phone network: smart network, dumb endpoints
– Internet: dumb network and smart endpoints
• Reinvented Internet: can we have best of both?
– Allow network providers to provide value added services
– Empower end users to innovate
– Allow user-centric innovations to migrate to networks
“Democratization of Innovations” by Eric von Hippel
Manufactured-centered and user-centric innovations
32
Empowers Network Providers & Users
Sensor Network
Edge Site
Mobile Wireless Network
Federated Facilities
• Supports virtual network, provides protection and isolation
• Empowers end users to create virtual networks of their spec
33
IOA Theme Challenges
• Overall architecture and protocols
– Lot of hard problems to address
• Security
– Off by default, flow/capability, virtual networks
– New challenges with user’s virtual networks
• Applicability to cellular infrastructure
• Economic viability and new policy requirements
34
Service Oriented Architecture Theme
• Current services emphasis assumes Internet cannot be
changed.
• What if we can do services and Internet together in a clean
slate way?
• Internet does more than packet delivery
– Information dissemination
– Integration of sensor networks
• Data aggregation, data-oriented connectivity and search
– Virtualization
• service customization, resource allocation & isolation
– And more?
• Lot of interesting problems and opportunities
35
Technology Driven Architecture
Theme
• Technologies offer new capabilities: how to exploit them?
– Optical networks: dynamic circuits and topologies
– Cognitive radios: flexible use of spectrum, dynamic topologies
– Location awareness
• Technologies have peculiarities: how to expose them?
– Wireless: error rate, fading, …
– Sensor networks: energy efficiency tradeoffs
• Cross layer ideas show lot of potential
• Challenge and opportunity: how to create a workable
architecture around cross layer ideas?
36
Future Internet Node: Wired
• Supports
– Switching and routing
– Programmability and virtualization
– Multiple protocol stacks & user’s virtual networks
• Provides
– High capacity
– Smaller foot print
– Lower power
• Exploits technologies
– VLSI: ASIC, FPGA, multi-core, smart memories
– Optical transmission within and outside the system
– Opto-electronic switching and routing
37
Future Internet Node: Wireless
• Supports
– Broad spectrum and multiple bands
• Cellular, fixed wireless, unlicensed bands, …
– Programmable radio interfaces and cross-layer design
– Modularity & programmability for virtual networks (new services)
• Provides
– High capacity
– Smaller foot print
– Lower power
• Exploits technologies
–
–
–
–
–
Cross-layer optimization
MIMO
Software defined radios and cognitive radios
VLSI: ASIC, FPGA, multi-core, …
Commodity hardware: server blades
38
A Century of Innovations
• Infrastructure innovations have transformative
impact on the society
–
–
–
–
–
–
Electrification
Water supply and distribution
Highways
Radio and television
Telephone
Internet
• Every established infrastructure has undergone
major overhaul a few times
– Internet cannot and will not be an exception
39
Summary
• Internet impacted the world in a way few inventions have
– Federal agencies & research community played a critical role
• Urgent and important need to reinvent the Internet
– We have an opportunity and obligation to lead
• GENI a compelling initiative in support of this mission
The result may be even greater and far-reaching than
the invention of current Internet
Stanford has an opportunity to lead
40
Acknowledgments
• The GENI Planning Group
– Peterson, Anderson, Blumenthal, Casey, Clark, Estrin, Evans,
McKeown, Raychaudhuri, Reiter, Rexford, Shenker, Turner,
Wroclawski,
• The GENI Working Groups
–
–
–
–
–
Research Coordination
Facility Architecture
Backbone
Mobile wireless sensor networks
Distributed services
• Planning grant workshops participants
• CISE GENI Team
• And others
41
More Information
• www.nsf.gov/cise/geni
• www.geni.net
42